Hey HN!

I've been working on knowledge / memory / retrieval tooling for AI + human interaction for a few months now.

Then, just a few days ago Andrej Karpathy posted the subject on X and I felt like this would be a good time to get it out there https://x.com/karpathy/status/2039805659525644595

The short pitch in my view is "Second Brain, for agents, or humans, or a mix of both".

I've been using it personally for a while now, as the tooling has evolved, especially for access to my obsidian notes just like karpathy talks about (read some places, write others etc.). Especially handy when your vault is littered with PDF's and screenshots like mine.

It's also a problem I've worked with a bunch in the enterprise, trying to connect AI & SaaS tools to disparate sources of knowledge, extracting contents from them, making them effectively retrievable, and then controlling the levels of write-back tolerated by agents across different spaces.

I've put as much work as I can into extensive documentation for v0.1, but I'm sure there's lots of improvements needed and some video tutorials & screenshots to follow up with to give some concrete use case examples.

For now, if you're interested in what I think is the most common example (docker self-hosted instance), here is a subfolder with that setup and some inline explainer comments: https://github.com/akunasoftware/akuna/tree/main/examples/co...

If you're interested but the lack of native installer support is a barrier for you, I'd love to get some prioritisation feedback on where to put those efforts. I have some work-in-progress locally but would like to know which platforms to focus on first besides the existing support.

Thanks for taking the time, hope it interests you as much as me, questions & feedback much welcome!

The Hacker News Tarpit

Ask HN: Has your company implemented agentic coding?

Agentic memory: the field is converging – but we're measuring the wrong thing

Fujitsu One Compression

Show HN: Invariant – pre-execution control layer for agentic workflows

Show HN: Separating work and play in Claude Code

Goldman Sachs to laid-off tech workers: take time, earnings loss to find new job

NY Yankees' torpedo bat is the same as regular bat

Show HN: Knowledge Bases for AI/Human Sharing

High AI judgment consistency does not mean high reasoning quality (preprint)

Hello World

Show HN: Where Is Artemis? Realtime open source 3D tracker for Artemis 2 mission

OpenAI's vision for the AI economy: public wealth funds, robot taxes, and more

Factory Makes the Most Expensive Stuff [video]

Show HN: Physical constants from 2 integers – MIT, 1225 tests, falsifiable

Sourcehut disrupted due to DDoS attack

I use Pollinations as a lightweight LLM back end (faster than Hugging Face)

RAPP Brainstem

jq for Forensics

3x faster project loads with the origin private file system

Audicin Wants to Fix Your Stress with Sound

Lunar Communications Relay and Navigation Systems

AI-Displaced Workers Could Face Long Setbacks, Report Finds

Patch Claude Code system prompt to fix laziness

How Do You Find an Illegal Image Without Looking at It?

Claude Code limits are starting to feel like a psychological trick

I Deploy Apps to My Homelab (A Love Story in 9 Acts)

Exploding primordial black holes might have reshaped the early universe

Open-source sandboxes to run AI code, browser agents and computer-use

Inspired by gstack: I stopped prompting Claude and gave it job titles instead

Show HN: MemberLane – Paid Communities on Telegram, Discord, and WhatsApp